I have expected that the course will be smooth to follow. However, training the models either of thumbs classification or of image regression are not possible on Jetson Nano, Either with the monitor attached, or with headless mode (That it looks it consumes less resources) Jupyterlab keeps failing.
I have read that Jetson Nano works well for inference.
- Is there a way to carry on the course by teaching how to train models using some cloud service?
Thanks in advance.
Do you use Nano 2GB?
More, could you share what kind of error do you meet?
Hi @janidai, if you are using MIPI CSI camera with the Nano DLI notebooks, you may want to try USB camera instead, or try these suggestions. You can also check that you disabled ZRAM and mounted SWAP like shown here. Also the Nano DLI is meant to be run headlessly, without display attached.
You can also try these DNN trainings from Hello AI World - https://github.com/dusty-nv/jetson-inference
They should run on Nano 2GB.